scfinder configs for running in real-time

[dufangqing]

Hi there,

To run scfinder in realtime, I simply executed "seiscomp-finder exec scfinder --debug', because I assume scfinder would read inventory and record streams from the host by default.

However, in my case, it seems scfinder was only able to read the inventory from the host but not the real-time streaming waveforms since I kept getting the "too few stations in raw data list, exit trigger check!" info.

I wonder if there are extra configs that I have to specify and do you think the connection error message I get from the log is critical for the issue?

...
16:01:29 [info] load_templates: Time passed in seconds for loading templates: 1.13259
16:01:29 [debug] FinDer version is finder-3.2.4_2023-10-31
16:01:29 [debug] GMT version is: 5.4.5
16:01:29 [info] Throttle Finder processing to an interval of 1.000000s
16:01:29 [info] Starting record acquisition
16:01:29 [info] Starting message thread
16:01:29 [error] SocketException: Socket connect error; Try to reconnect
16:01:30 [info] determine_trigger: too few stations in raw data list, exit trigger check!
16:01:31 [info] determine_trigger: too few stations in raw data list, exit trigger check!
16:01:33 [info] determine_trigger: too few stations in raw data list, exit trigger check!
16:01:35 [info] determine_trigger: too few stations in raw data list, exit trigger check!
....

ps., a little configuration issue that I noticed: if adding "database.inventory = host.docker.internal/seiscomp" to the config file, I would get the error below, one may just exclude this specification to avoid the issue...

16:01:08 [info] Loading complete inventory from host.docker.internal/seiscomp
16:01:08 [error] host.docker.internal/seiscomp not found

Thanks a lot,
Du

[FMassin]

Dear @dufangqing

Socket connet error

16:01:29 [error] SocketException: Socket connect error; Try to reconnect

Means that scfinder cannot connect to the seedlink server. This message is common to all seiscomp modules using seismic data (i.e., scautopick, scamp, and sceewenv & scfinder from the EEW addons). So I think you need to go back to slinktool and seedlink to understand what's going on. I suspect seedlink to be crashed.

database.inventory error

Depending on the version, the global seiscomp parameter database.inventory might require database type (e.g,. mysql://), your response is great, you can also fix it with:

database.inventory = mysql://host.docker.internal/seiscomp

But this kind of complex config is irrelevant to 99% of users so I will remove it from the documentation. Thanks a lot for spotting this.

[dufangqing]

Thank you very much @FMassin. The connection problem was solved by specifying --record-url slink://host.docker.internal:18000.

[dufangqing]

however, I received data latency warnings (e.g., 09:05:44 [warning] OX.CSM..HNN: max delay exceeded: 17.408047s) while running scfinder in real-time. But I guess this latency is from my seismic network itself?

[FMassin]

Yes, it should reflect what's called latency provided by slinktool -p -S 'OX.CSM:HNN' host.docker.internal:18000.

[dufangqing]

Thanks! Indeed there's latency in the network.